Michael K. Bourdoulis

Direct Power Control of DFIG Wind Systems Based on Nonlinear Modeling and Analysis

A complete new modeling approach for doubly fed induction generator (DFIG) wind energy systems is provided in this paper. The model incorporates all the system component dynamics, i.e., the ones of the induction generator and the ac/dc/ac frequency converter, and is developed on state space with the state vector including directly as states the stator-circuit and grid-side converter active and reactive powers. This innovation, combined with a voltage-oriented model deployment, permits the design and application of simple feedback PI direct power controls (DPCs). Hence, the proposed design approach becomes independent from the flux measurement or estimation under the cost of requiring a rigorous stability analysis, caused by the fact of not using field-oriented vector control techniques. Using recent advanced nonlinear methods, this analysis is completely performed on the entire closed-loop nonlinear system to conclude input-to-state stability and convergence to the desired equilibrium. Thus, a fully analyzed design approach for DPC is provided with guaranteed stability, further evaluated through extensive simulation results on a commercial 2-MW DFIG wind system.

Rotor-side PI controller design of DFIG wind turbines based on direct power flow modeling

In this paper, an appropriate nonlinear modeling approach of a DFIG wind system is introduced that permits the design of suitable PI controllers for direct regulation of the active and reactive power produced. Using this model and adopting grid voltage reference frame orientation, it is immediately shown that the stator power components, namely the reactive and active power, can be controlled separately through the d- and q-axis rotor voltage inputs, respectively. This provides the possibility to design simple PI controllers for each voltage input, thus overcoming the conventional cascaded controller designs used so far. Therefore, the proposed methodology does not require either current inner-loops or estimation of any flux components. Furthermore, the proposed modeling and control scheme provides a closed-loop system in a form suitable for applying advanced, Lyapunov-based, stability analysis methods. As shown in the paper, by using these methods, stability and convergence to the equilibrium can be guaranteed. Eventually, to verify the analysis and evaluate the performance of the closed-loop DFIG wind system, extensive simulations under wind speed changes and reactive power reference variations are conducted.

Direct power flow modeling and simple controller design for ac/dc voltage-source converters

In this paper, instead of using traditional cascaded control structures, a model-based simple controller scheme is proposed. This scheme, developed on the synchronously rotating dq reference frame, enables the direct active and reactive power regulation in contrast to the usually applied schemes that utilize inner-loops to regulate the d and q current components at a reference, as it is provided by outer-loop active and reactive power controllers. Extracting the suitable model and adopting the grid voltage orientation, decoupling of the control inputs is immediately observed that permits for each input to separately regulate the active and reactive power, respectively. However the main contribution of this study, with respect to other existing direct power control methods, is its theoretical analysis that proves a stable operation and convergence to the desired equilibrium. The analysis and the performance of the nonlinear closed-loop system are verified through simulation results.

Model analysis and intelligent-ANFIS control design for PWM-regulated ac/dc converters

A new design approach for nonlinear intelligent controllers for PWM-regulated ac/dc converters is proposed. First, the nonlinear model of the complete system is obtained, on which a passivity based analysis is conducted. Using this model, it is proven that the system always has the damping property independently from the controller scheme that one can apply. This is a significant contribution for fuzzy logic based intelligent control applications since it makes possible to overcome the inherent drawback of these designs, that is the ability to guarantee stability of the closed loop system. Thus, nonlinear adaptive neuro-fuzzy inference system (ANFIS) controllers are proposed to be used in order to obtain flexible, fast and efficient regulation of PWM-regulated ac/dc converters. The performance of the proposed ANFIS based controllers is simulated on a test system. The transient response of the proposed ANFIS controllers indicates an excellent performance which seems to enhance that of the conventional PI controllers.

PI controller design of grid-side PWM-regulated ac/dc converters via stability analysis based on passivity

PWM-regulated ac/dc converters are of great importance in distributed power system applications since they are widely used to interface a variety of renewable power sources such as wind and photovoltaic generators with the grid. Furthermore, they provide special control possibilities of the active and reactive power. In this paper, a new stability analysis and design approach for nonlinear PI controllers for this kind of devices is proposed, based on the passivity analysis of the complete system. To this end, first the nonlinear model of the system is obtained, on which suitably modified PI controllers are incorporated. The analysis conducted provides a set of limits, rules and demands that should be satisfied in order to achieve fast, accurate and satisfactory responses. Furthermore, it is proven that, in all cases, a closed-loop system with desired damping is guaranteed while the controllers dynamics drive the system at the reference values. The simulation results indicate the effectiveness of the controller design.

An alternative PI controller design approach for PWM-regulated ac/dc three-phase converters

A new design for cascaded PI controllers used in PWM-regulated ac/dc converters is proposed. The design makes possible the selection of the PI controllers gains in a manner that is more independent from the system parameters. To this end, a system and controller analysis is conducted that results in the determination of an effective time constant for the closed-loop system. It is proven that this variable time constant results from the response of an arbitrarily selected lead or lag transfer function. This enables to select an arbitrary integrator gain that leads in an initial time constant which is corrected by the selection of suitable proportional gain values. For a given integrator gain, small enough proportional gain values lead to slower responses while large enough proportional gain values lead to faster responses. The evaluation of the proposed design approach is easily obtained by simulation results.

An alternative modeling and controller design guaranteeing power stability for DFIG wind systems

Doubly-Fed Induction Generators (DFIG) are widely used in wind power systems due to their inherent capability of controlling the produced active and reactive power at desired levels, in a large range of wind speeds. In this paper, a nonconventional modeling approach of a DFIG wind system is introduced that permits to control directly the active and reactive power produced. To this end, first, the complete nonlinear dynamic model that contains as states the stator active and reactive power is extracted in the synchronously rotating dq reference frame. Assuming operation under grid voltage reference frame orientation, it is easily shown that the stator power components can be controlled separately through the d- and q-axis rotor voltage inputs. Hence, unlike the complex conventional cascaded controller designs for DFIGs, in this paper, a simple design of proportional controllers for the stator power components is adopted. For this scheme an advanced, Lyapunov-based, stability analysis is conducted that guarantees stable operation and convergence to the equilibrium. This closed-loop scheme is further completed by an outer PI controller design that tracks the rotor speed to the optimum, providing the active power reference for the maximum power point operation. Finally, the analysis and the performance of the closed-loop DFIG wind system are verified through simulation results.

A new controller design and analysis of DFIG wind turbine systems for MPP operation

In this paper, a simplified and easily implemented MPP control scheme is proposed for DFIG wind turbine systems. In particular, taking into account the complete nonlinear model of the system in the synchronously rotating dq reference frame operating at the voltage oriented control mode, a cascaded q-axis controller scheme and a simple proportional d-axis controller are proposed. The inner-loop of the cascaded q-axis controller does not require decoupling networks and is used to achieve the desired q-axis current damping. The outer-loop PI controller enforces the MPP operation of the system by continuously tracking the rotating speed to the optimum. Its structure is slightly modified in a way that permits a second order linear system analysis. Stability and convergence to the equilibrium is proven by taking into account the inner-loop current controllers in the nonlinear analysis of the complete system. Extensive simulation results on a 2MW DFIG wind turbine system illustrate the enhanced system performance and verify the effectiveness of the controller.

Distributed generation power system modeling in nonlinear Hamiltonian form

Distributed generation has dramatically changed the structure of modern power systems. In this structure, power electronic devices are extensively used providing the possibility of new control strategies in the distribution network. To implement these strategies, a complete dynamic analysis of the distributed generation system is needed. In this paper, exploiting a common feature of almost all the distributed generation components that is their individual modeling in Hamiltonian form, we propose a systematic methodology of obtaining the complete distributed generation system model. Furthermore, we show that this model is also in Hamiltonian form with certain damping properties that can be effectively used for stable control designs. As an example, a particular distributed generation system that includes wind and photovoltaic generations is modeled and simulated.